[Tagging] Rock outcrops in forest

Christoph Hormann osm at imagico.de
Sun May 27 11:38:11 UTC 2018

On Sunday 27 May 2018, Ture PĂ„lsson wrote:
> How should these areas be tagged in OSM? Some of them have been
> imported as natural=bare_rock overlapping with landuse=forest, but
> this has led to protests from persons stating that bare_rock should
> be used only for areas of naked rock, with little or no vegetation
> (and, looking at the Wiki, I can only agree with them).

Short answer: You can only determine this on a case by case basis.

Longer answer: If you have something in reality that does not really 
match any of the established tags in OSM you have three options:  

1) Choose a best fit tag according to your assessment (and possibly 
supplement it with additional tags).
2) Invent a new tag to document exactly what you have.
3) If what you have features some spatial granularity (in your case: 
small areas with trees mixed with small areas without trees) you can 
use more small grained mapping to resolve this.

> (Do these areas belong on OSM at all? I believe they do, as they
> provide interesting information e.g. for hikers. I admit, however,
> that they are often hard to detect on the aerial photos available to
> us, and that their borders are somewhat unclear and difficult to
> determine.)

There is currently no established tagging to characterize the ground in 
a forest/wood in OSM even though this is a major differentiating 
element.  Dense forests, in particular with coniferous trees, often 
don't have much underbrush but usually a top soil layer with high 
organic content from dead leaves/needles.  Less dense forests can have 
all kinds of lower growing vegetation layers.  And in dry climate or 
recently glaciated areas where soil cover is thin and incomplete like 
in Scandinavia or Canada you can also have either loose or solid rock 
as the main ground condition under the trees.

I think i suggested this before already - one way to allow documenting 
this in a generic form (i.e. that does not exclusively deal with the 
wood/forest case but also any other similar situation) would be to 
introduce a natural:secondary=* tag.  That is natural=wood + 
natural:secondary=grassland would mean a woodland with grass growing 
under the trees.  And natural=grassland + natural:secondary=wood would 
be a grassland with scattered trees.  This would likewise work with 
bare_rock.  You would still have to decide if this is primarily a 
woodland or a bare rock area but you could document the fact that both 
elements are present.

Christoph Hormann

More information about the Tagging mailing list